Identifying rock fracture precursor by multivariate analysis based on the digital image correlation technique

Estimating the initiation and propagation of rock fractures and identifying fracture precursors are significant for rock failure analysis. This study performed numerical uniaxial compression tests of rock samples with an excavated hole and laboratory tests of rock-like samples. Statistical methods were used to analyze the precursory indices of full-field strain information based on digital image correlation (DIC) technology. An identification method for rock fracture initiation based on the absolute standard deviation (ASD) and absolute variation coefficient (AVC) of the strain field was proposed. The results indicated that a sudden change in the shear strain field during the loading process was closely related to the occurrence of internal cracks in the rocks. The variation laws of the ASD and AVC of the strain field were highly correlated with the rock fracture evolution. The progress in the variation rate of the strain field corresponded to the rock fracture evolution, and the change points can be used as precursors of the rock fracture. The precursory points of the excavated and rock-like samples in each crack propagation stage were at least 0.84% and 1.06% earlier than those of the current stage, respectively. The precursory points of each stage were 2.48% and 2.11% earlier than those of the final failure stage. The first precursory points were 45.45% and 42.11% earlier than the final failure time.